Global Information Technology Industry: Comprehensive Analysis and Strategic Outlook (Week Ending 17 July 2026)

The week ending 17 July 2026 represented a critical inflection point for the global Information Technology (IT) sector, characterised by the compounding acceleration of artificial intelligence (AI) infrastructure investments, aggressive corporate consolidation, and a pronounced geopolitical fracturing of digital governance. The macroeconomic environment, as articulated by the International Monetary Fund (IMF), is increasingly defined by a technological bifurcation. Economies tightly integrated into the AI hardware and software value chains are exhibiting robust resilience, whilst energy-importing nations grapple with the inflationary pressures of geopolitical conflicts¹.

This divergence is acutely visible within the semiconductor and data centre markets. Record-breaking financial disclosures from Taiwan Semiconductor Manufacturing Company (TSMC) and ASML Holding NV confirmed that the hyperscale capital expenditure supercycle is far from peaking, driven by the industry-wide transition from generative models to autonomous, agentic AI systems³. However, this transition is exposing severe operational frictions within legacy technology conglomerates. Alphabet faced significant market scrutiny following the indefinite delay of its flagship Gemini 3.5 Pro model due to substandard coding performance, whilst Apple launched a historic trade-secret lawsuit against OpenAI to protect its consumer hardware moat against rapid talent poaching⁵.

Simultaneously, the geopolitical architecture of the internet is undergoing a formal realignment. China spearheaded the inauguration of the World Artificial Intelligence Cooperation Organisation (WAICO), establishing a Beijing-headquartered institutional counterweight designed to coalesce Global South nations against United States-led export controls⁷. Concurrently, sovereign nations ranging from Australia to individual US states enacted sweeping domestic regulatory frameworks to enforce operational guardrails on AI deployment and hyperscale infrastructure⁹. Finally, a severe escalation in highly coordinated ransomware campaigns, dominated by the Ransomware-as-a-Service (RaaS) syndicate ‘The Gentlemen’, paralysed critical supply chains across the globe, prompting decisive sanctions from the US Department of the Treasury¹¹.

Macroeconomic Dynamics and the Semiconductor Supercycle

The Bifurcated Global Tech Economy

The macroeconomic foundation of the IT sector was starkly outlined in the IMF’s July 2026 World Economic Outlook Update, which projected global GDP growth at 3.0 per cent for 2026 and 3.4 per cent for 2027¹. This headline stability masks severe underlying volatility and a fundamentally asymmetric recovery. The IMF explicitly noted that economies deeply integrated into the global technology and AI value chains are capturing a disproportionate share of global growth¹. The immense capital flows directed toward data centre construction, advanced logic chip fabrication, and power infrastructure are effectively insulating these technology-centric economies from the macroeconomic shocks weighing heavily on energy importers reliant on the Middle East¹.

This “K-shaped” technological recovery places immense pressure on enterprise technology earnings. As the major hyperscalers—Microsoft, Amazon, Meta, and Alphabet—approach their earnings disclosures, the market paradigm is shifting¹⁴. Investors are no longer rewarding indiscriminate capital expenditure on AI servers; rather, the focus has pivoted to “tokens per watt” efficiency and clear pathways to monetising agentic AI capabilities¹⁴. The market is demanding empirical evidence that the hundreds of billions of dollars being poured into graphics processing units (GPUs) and networking equipment are translating into durable software revenue streams rather than merely inflating depreciation costs¹⁵.

ASML and TSMC: Leading Indicators of Sustained Infrastructure Demand

The strongest empirical evidence that the AI infrastructure boom remains robust was delivered by the foundational pillars of the semiconductor supply chain: ASML and TSMC. Because semiconductor fabricators must order equipment and reserve fabrication capacity years in advance, the financial forward guidance of these two entities serves as the most accurate leading indicator for the broader IT sector.

ASML, the Dutch monopoly provider of extreme ultraviolet (EUV) lithography systems, reported phenomenal second-quarter results. The company recorded €9.33 billion in total net sales and €2.92 billion in net income, substantially outperforming market consensus¹⁶. Driven by relentless demand from manufacturers like TSMC, Intel, and Samsung, ASML increased its full-year 2026 revenue guidance to a range of €43 billion to €45 billion⁴.

Source: ASML Q2 2026 Financial Results¹⁷.

Crucially, ASML announced aggressive capacity expansion plans, aiming to increase its production of low-NA EUV systems and deep ultraviolet (DUV) immersion systems by 30 per cent in 2027, with investigations underway for an additional 30 per cent expansion in 2028¹⁸. This multi-year capacity commitment signals that ASML’s customers foresee a sustained deficit in global compute capacity. TSMC corroborated this outlook by posting a record second-quarter net profit of NT$706.56 billion, representing a 77.4 per cent year-on-year increase, driven primarily by high-performance computing²⁰.

In a massive vote of confidence for the United States’ semiconductor reshoring efforts, TSMC committed an additional $100 billion to its Arizona manufacturing hubs³. This brings the company’s total investment in the US to $265 billion, aimed at deploying 2-nanometre-and-below logic processes and expanding CoWoS advanced packaging capacity³. The alignment of ASML’s elevated order book and TSMC’s capital expenditure surge strongly implies that the transition from generative AI to complex, multi-step agentic AI will require a hardware footprint vastly larger than current market estimates³.

The Global Data Centre Arms Race

The physical manifestation of this semiconductor demand is a global explosion in hyperscale data centre construction, accompanied by rising geopolitical and environmental friction regarding energy consumption. In the US, Microsoft brought its $3.3 billion Mount Pleasant AI campus in Wisconsin online, while Alphabet raised $80 billion in equity capital to finance further infrastructure²². Amazon and Google collectively pledged $25 billion for developments in Montgomery County, and Microsoft secured a 20-year dedicated power agreement with Chevron to fuel a massive facility in Texas²².

Internationally, the scale of deployment is equally staggering. In Europe, SoftBank initiated a €75 billion project to construct a 5-gigawatt AI data centre in northern France, and hyperscaler Hscale acquired a 250-megawatt campus in Milan²². In the Asia-Pacific region, Blackstone-backed AirTrunk outlined a $21 billion investment for a 3-gigawatt facility in Maharashtra, India, while Meta partnered with Reliance Industries for its inaugural AI-enabled Indian data centre, backed by 1 gigawatt of renewable energy²².

However, this rapid physical expansion is triggering severe regulatory pushback. The state of New York enacted the first statewide moratorium on the construction of new hyperscale data centres, pausing developments for up to a year to assess the systemic risks these facilities pose to the local energy grid and water supplies²³. This regulatory hesitation highlights a growing global tension: governments are eager to capture the economic benefits of the AI supply chain but are increasingly alarmed by the unprecedented environmental and infrastructural toll exacted by advanced computing facilities²³.

Artificial Intelligence: Foundational Models and The Agentic Shift

Google’s Gemini 3.5 Pro Delay and the “Tacitus Trap”

The competitive benchmark within the AI sector has unequivocally shifted from conversational fluency to agentic capability—the capacity of a model to autonomously navigate operating systems, utilise digital tools, and execute complex coding architectures without continuous human prompting²⁵. In this highly contested arena, Alphabet suffered a severe operational failure.

Google was forced to indefinitely delay the launch of its flagship model, Gemini 3.5 Pro, missing its heavily publicised June release window⁶. The delay triggered acute frustration among Google’s engineering ranks and precipitated a 4 per cent drop in Alphabet’s share price during mid-week trading⁶. The core issue plaguing Gemini 3.5 Pro is its substandard performance in AI-assisted coding and agentic workflows⁶. Despite an emergency initiative in late June to update the model’s training data with extensive programming-related datasets, the evaluation results failed to meet internal benchmarks, forcing the DeepMind division back into the development phase⁶.

Industry analysts attribute this delay to systemic bureaucratic friction within Google, a phenomenon commentators have dubbed the “Tacitus Trap.” Unlike leaner competitors, Google must homogenise its flagship models to function seamlessly across a sprawling, legacy product ecosystem encompassing Search, Android, YouTube, and Google Workspace²⁹. This requirement for universal compatibility creates immense stakeholder paralysis, preventing the rapid iteration cycles required to maintain parity at the AI frontier²⁹.

Anthropic and OpenAI Capitalise on the Agentic Gap

Google’s stagnation allowed its primary competitors to solidify their market leads. Anthropic successfully deployed Claude Sonnet 5, executing a masterclass in pricing strategy and enterprise capability. Launched as the default model for all users, Sonnet 5 achieved near-parity with Anthropic’s flagship Opus 4.8 model but at a fraction of the inference cost²⁵. The model achieved a 80.4 per cent success rate on the Terminal-Bench 2.1 evaluation (agentic command-line engineering) and a 63.2 per cent score on agentic coding benchmarks, proving that highly capable autonomous agents no longer require prohibitive compute budgets²⁵.

Simultaneously, OpenAI launched GPT-5.6 Sol, a model specifically engineered for token efficiency in agentic coding tasks⁶. The concurrent successes of Anthropic and OpenAI highlight the precarious position of legacy technology giants. As agentic AI transitions from experimental demonstrations to deployed enterprise operating systems, companies failing to deliver reliable, multi-step autonomous reasoning risk rapid obsolescence in the foundational model layer¹⁵.

Corporate Warfare and Market Consolidation

The Apple vs. OpenAI Trade Secret Litigation

The transition from software-based large language models to AI-native hardware has ignited vicious corporate warfare. On 11 July 2026, Apple filed a sprawling, high-stakes trade-secret lawsuit against OpenAI in the United States District Court for the Northern District of California⁵. This litigation represents a spectacular unravelling of the brief partnership forged between the two entities in 2024 to integrate ChatGPT into Apple’s Siri assistant³⁴.

Apple’s complaint outlines a highly coordinated, institutionalised campaign of corporate espionage orchestrated by OpenAI to shortcut the development of its nascent consumer hardware division⁵. Central to the dispute is OpenAI’s 2025 acquisition of ‘io Products’, a hardware startup founded by former Apple design chief Jony Ive, which served as the beachhead for a massive talent raid resulting in the defection of over 400 former Apple engineers to OpenAI³³.

The lawsuit specifically names Tang Yew Tan, a 24-year Apple veteran and former vice president of product design for the iPhone, who now serves as OpenAI’s Chief Hardware Officer⁵. Apple alleges that Tan systematically directed interviewing Apple employees to bring physical prototypes, engineering blueprints, and confidential component specifications to their OpenAI job interviews for “show and tell” sessions³⁴. Furthermore, the complaint names Chang Liu, a former senior system electrical engineer, accusing him of retaining an Apple-issued laptop post-resignation and exploiting a rare authentication vulnerability to download dozens of proprietary hardware files relating to battery chemistry, display technology, and power management silicon⁵.

The strategic implications of this lawsuit extend far beyond the immediate theft of intellectual property. By launching an aggressive discovery phase and serving legal preservation letters to dozens of OpenAI employees, Apple aims to legally entangle OpenAI’s hardware ambitions right as the AI lab prepares for a highly anticipated IPO⁴¹. OpenAI’s objective to build an ambient, screenless AI gadget represents an existential threat to the iPhone’s dominance as the primary consumer computing interface³⁶. Apple is leveraging the courts to aggressively defend its hardware moat, signalling to the entire industry that it will not tolerate the wholesale extraction of its institutional knowledge³⁶.

Stripe and Advent’s $53 Billion Play for PayPal

While Apple and OpenAI battle over physical hardware, the digital payments architecture witnessed a monumental consolidation attempt. Payments infrastructure provider Stripe, operating in a consortium with private equity giant Advent International, launched a $53 billion unsolicited takeover bid for PayPal Holdings⁴⁴. The consortium offered $60.50 per share in cash, representing a 28 per cent premium over PayPal’s trading price prior to the announcement⁴⁶.

The magnitude of the proposed transaction is underscored by the $50 billion in committed bank financing secured by the bidders, demonstrating immense institutional confidence in the strategic rationale of the merger⁴⁷. PayPal has languished in recent years, suffering from sluggish revenue growth and severe margin compression as Apple Pay and Google Pay steadily eroded its market share at digital checkouts⁴⁴. Consequently, PayPal’s market capitalisation had plummeted from a pandemic-era peak of $360 billion to under $50 billion in early 2026⁵¹.

However, the underlying catalyst for Stripe’s aggressive move is the architecture of global stablecoins. Stripe possesses ‘Bridge’, a premier developer-first platform enabling businesses to seamlessly issue dollar-backed stablecoins⁴⁸. PayPal, conversely, controls ‘PYUSD’, a highly liquid consumer stablecoin integrated into a network of over 400 million active retail accounts and merchants⁴⁸. By acquiring PayPal, Stripe intends to vertically integrate the entire stablecoin stack—from backend issuance to retail distribution—creating an unprecedented global settlement network capable of bypassing traditional banking rails⁴⁸. While the bid drove PayPal’s stock up by over 16 per cent, the proposed equal-ownership structure between Stripe and Advent will undoubtedly face fierce antitrust scrutiny from regulators wary of a massive concentration in the merchant acquiring market⁵¹.

Strategic Divestments: Project44

In a parallel move highlighting the deepening integration of AI into global supply chains, logistics visibility giant Project44 announced the strategic bifurcation of its operations. The company spun off a dedicated entity named ‘LSP44’, designed specifically to provide purpose-built AI agents and Application Programming Interface (API) infrastructure for logistics service providers⁵⁴. This structural separation allows the core Project44 business to concentrate entirely on serving enterprise shippers via a Decision Intelligence Platform, demonstrating how legacy software firms are restructuring to commercialise industry-specific, autonomous AI agents⁵⁴.

Geopolitics, Sovereignty, and AI Governance

The week was defined by intense legislative and diplomatic manoeuvring as sovereign entities sought to assert dominance over the rapidly evolving digital landscape. The underlying theme was a pronounced rejection of a monolithic, unipolar technology ecosystem dominated by Silicon Valley.

The Launch of WAICO and the Global South Alternative

The most significant geopolitical development was the formal establishment of the World Artificial Intelligence Cooperation Organisation (WAICO) on 16 July 2026⁷. Headquartered in Shanghai, the intergovernmental body was inaugurated via an agreement signed by 29 founding nations, prominently featuring China, Indonesia, Russia, Kazakhstan, and Pakistan⁷. The signing ceremony, attended by UN Secretary-General António Guterres, served as a powerful declaration of intent by the Global South to construct a parallel AI governance and development framework⁷.

WAICO’s stated mission is to ensure equitable access to AI technology, promote open-source ecosystems, and bridge the widening digital divide⁵⁶. During his keynote address, Chinese President Xi Jinping firmly rejected the monopolisation of AI by a “handful of countries,” promising to provide 5,000 AI training and capacity-building opportunities to developing nations over the next five years⁵⁷. Furthermore, China pledged to establish international AI application centres in coordination with ASEAN, the African Union, and BRICS nations⁵⁹.

This institutional manoeuvring is a direct, calculated response to the escalating tech war with the United States. WAICO positions itself as the ideological counterweight to ‘Pax Silica’, the US-led coalition attempting to tightly control the global supply chains of advanced semiconductors, critical minerals, and AI inference capabilities⁸. The necessity of WAICO for non-aligned nations was starkly illustrated earlier in the month when the US government utilised extraterritorial export controls to temporarily sever international access to Anthropic’s Claude 5 models⁸. By offering access to powerful Chinese open-weight models and state-backed computing infrastructure, WAICO provides developing nations with a strategic hedge against the weaponisation of Western cloud software⁸.

Regulatory Realities in Europe and Asia

While the Global South coalesced around WAICO, the European Union confronted the systemic vulnerabilities of its own technological sovereignty. The European Commission published the 2026 State of the Digital Decade package, revealing a deeply flawed transition²⁴. Although basic digital infrastructure is sound, with 5G covering 96.8 per cent of households, the EU’s industrial base remains critically reliant on foreign entities²⁴. The continent accounts for merely 9 per cent of the global semiconductor market, and its ICT workforce has stagnated at 5 per cent of total employment—half of the 2030 target²⁴. Furthermore, soaring enterprise AI adoption (increasing by 48 per cent year-on-year) has placed immense strain on the continent’s computing capacity²⁴. In response, the EU is rapidly accelerating the deployment of the Cloud and AI Development Act and the Chips Act 2.0 to prevent total subservience to American hyperscalers²⁴. Concurrently, European military investment experienced a massive pivot towards autonomous systems, with vast capital injections into drone procurement, reflecting a fundamental shift in continental defence strategy²³.

In Asia, domestic regulation took a decidedly social turn. Chinese regulators implemented severe restrictions prohibiting generative AI models from fostering romantic or intimate relationships with users²³. The ban forced major platforms, including Alibaba and ByteDance, to rapidly disable companion chatbot functionalities²³. This decisive intervention highlights Beijing’s willingness to aggressively curtail technology applications deemed socially detrimental, standing in stark contrast to the comparatively laissez-faire approach to algorithmic social impact observed in Western jurisdictions²³.

Australia’s Mandatory AI Standards

The Australian government executed a comprehensive overhaul of its digital governance posture. On 15 July 2026, Prime Minister Anthony Albanese outlined a new, mandatory national AI framework, decisively abandoning the previous reliance on voluntary ethical guidelines¹⁰.

The framework is structurally anchored by the newly established Office of AI, situated within the Department of the Prime Minister and Cabinet, tasked with coordinating cross-portfolio compliance¹⁰. The legislation, expected to be tabled in early 2027, will impose rigorous operational mandates on data centre operators. Facilities will be legally compelled to underwrite new renewable energy supplies to ensure they remain net generators to the power grid, and they must self-finance any requisite water infrastructure¹⁰.

Crucially, the Albanese government explicitly ruled out granting technology companies broad text and data mining exemptions, instituting formidable copyright protections to shield the domestic creative and media sectors from algorithmic scraping¹⁰. Financially, the federal budget supported this pivot by allocating $70 million for ‘AI Accelerator’ grants to stimulate sovereign capability, alongside a $654 million injection to expand the national Digital ID system⁶⁸. The government also scheduled a nationwide test of the new ‘AusAlert’ emergency cellular broadcast system for late July, reflecting an overarching strategy to modernise and secure the nation’s core digital infrastructure against emerging threats⁶⁹.

The US Legislative Patchwork

In the absence of comprehensive federal AI regulation in the United States, individual state legislatures dramatically accelerated their own statutory interventions⁹. The New York data centre moratorium was accompanied by aggressive legislative drafting in Massachusetts, where bills S 3164 and S 2581 advanced to restrict algorithmic social media feeds targeting minors and mandate comprehensive mobile phone bans within public school environments⁹.

Ohio introduced a raft of targeted regulations, including HB 524, which imposes strict liability on entities whose AI models generate self-harm recommendations, and HB 469, a pre-emptive statute legally declaring AI systems non-sentient to preclude any future claims of algorithmic personhood⁹. Pennsylvania mirrored this proactive stance by advancing legislation to mandate cryptographic watermarking for AI-generated synthetic content and to enforce algorithmic transparency in health insurance determinations⁹. This fragmented, state-by-state regulatory patchwork is creating profound compliance complexities for enterprise AI developers attempting to deploy homogeneous software products across the American market⁹.

Cybersecurity, Ransomware Ecosystem, and Critical Infrastructure

The operational integrity of global IT networks faced relentless degradation during the week, characterised by the industrialisation of ransomware and a severe escalation in attacks against critical physical infrastructure. A comprehensive study by Comparitech revealed a chilling statistical reality: government agencies are now falling victim to successful ransomware encryption events at an average rate of once per day⁷⁰. The mean ransom demand has calibrated to approximately $100,000—a deliberate pricing strategy designed to fall beneath the threshold of catastrophic public scrutiny while ensuring a high probability of payment from tax-funded entities⁷⁰.

Compounding this threat is a fundamental shift in the initial access vector. Threat actors have largely abandoned complex zero-day software exploits in favour of identity-based attacks⁷¹. Exploiting compromised employee credentials, bypassing weak multi-factor authentication, and leveraging sophisticated social engineering are now the primary methodologies for breaching enterprise perimeters⁷¹.

The Dominance of ‘The Gentlemen’ Syndicate

The cybercriminal ecosystem has undergone a violent restructuring, resulting in the ascendancy of a Ransomware-as-a-Service (RaaS) syndicate known as ‘The Gentlemen’ (tracked by Microsoft as Storm-2697)⁷³. Usurping established giants like LockBit and Qilin, The Gentlemen have claimed over 478 victims globally within a span of mere months¹².

The syndicate’s rapid expansion is driven by a highly aggressive economic model, offering affiliates an unprecedented 90 per cent share of extorted ransoms⁷³. Technologically, the group is formidable. They deploy a custom-built, Go-based encryptor that simultaneously targets Windows, Linux, ESXi virtual environments, and Network Attached Storage (NAS)⁷⁴. This multi-platform capability allows the group to instantly paralyse an organisation’s entire infrastructure—including critical backup repositories—in a single, coordinated strike⁷⁶. Prior to encryption, the group utilises a bespoke EDR-killer framework (“GentleKiller”) to autonomously disable Microsoft Defender and wipe forensic event logs, effectively blinding the victim’s security operations centre⁷³.

Ironically, the internal operations of The Gentlemen were exposed following a successful cyberattack against their own hosting provider, 4VPS⁷⁶. The resulting database leak provided cybersecurity researchers with an unprecedented view into the syndicate’s backend, revealing that their entire administrative panel was constructed in three days using AI coding assistants, and exposing detailed communication logs regarding their systematic targeting of edge devices like firewalls and VPNs⁷⁶.

Operational Technology and Supply Chain Casualties

The profound vulnerability of physical supply chains to digital disruption was starkly demonstrated across multiple sectors:

• Nichirei Corporation (Japan): On 13 July 2026, Nichirei, Japan’s paramount cold-chain logistics and frozen food conglomerate, suffered a catastrophic cyberattack¹¹. To contain the breach and protect the personal data housed on its servers, the company severed its enterprise network⁸⁰. This action immediately halted inbound and outbound logistics across its network of 140 refrigerated distribution centres¹¹. The cascading impact on the Japanese food supply was instantaneous. Kentucky Fried Chicken Japan was forced to suspend mobile orders and restrict menus nationwide due to severe ingredient shortages, whilst major supermarket chains like Aeon and conveyor-belt sushi operators like Kura Sushi experienced acute stock outages¹¹. The incident exemplifies the fragile, hyper-connected nature of modern logistics, where a single IT failure can instantly cripple thousands of downstream businesses⁷⁹.
• Kudankulam Nuclear Power Plant (India): A severe data leak highlighted the systemic vulnerabilities introduced by third-party engineering contractors. The data-extortion group ‘World Leaks’ published a massive 14.3-gigabyte cache of documents containing over 19,000 highly sensitive files pertaining to the Kudankulam Nuclear Power Plant in Tamil Nadu⁸⁴. The exfiltration did not occur from the plant’s isolated operational technology (OT) network, but rather from a private cloud server hosted by Yotta Data Services, belonging to the plant’s infrastructure contractor, Reliance Group⁸⁰. While the core Russian-designed VVER-1000 reactors remained secure, the leaked repository included highly detailed engineering blueprints for ventilation systems, cooling architecture, and the complete floor layout of the common control room for Units 3 and 4⁸⁴. Cybersecurity experts warned that this high-fidelity documentation provides hostile nation-states with the precise topographical mapping required to execute sophisticated, targeted sabotage or espionage operations against the facility’s support infrastructure in the future⁸⁰.
• Coca-Cola Fairlife (United States): A targeted ransomware deployment forced the immediate suspension of production at Coca-Cola’s Fairlife dairy unit across all US manufacturing facilities⁸⁹. The attack underscores the increasing frequency with which cybercriminal syndicates are targeting the agricultural and food production sectors, recognising that the imperative to prevent perishable product spoilage heavily incentivises the rapid payment of ransoms⁸⁹.
• Ernst & Young (EY) Data Breach: In the corporate services sector, global consultancy EY disclosed a significant breach stemming from its third-party IT support ticketing platform⁹¹. Attackers exploited the platform over a three-week window to meticulously exfiltrate support tickets containing sensitive tax filings and investment holding data belonging to EY’s institutional clients⁹¹. The incident highlights the acute risk posed by centralised IT service management tools, which often inadvertently aggregate vast quantities of highly sensitive client data in poorly secured, third-party environments⁹¹.

Sovereign Enforcement: Dismantling the Cybercrime Ecosystem

Recognising the failure of defensive measures alone, global law enforcement escalated offensive operations against the foundational infrastructure of the ransomware ecosystem. The United States Department of the Treasury’s Office of Foreign Assets Control (OFAC) levied severe economic sanctions against ‘First VPN Service’ (1VPNS), its Ukrainian administrator Dmytro Rashevskyi, and Belarusian cryptor developer Yegeniy Silayev¹³.

1VPNS functioned as a critical enabler for global ransomware operations. Operating as a strictly “no-logs” provider, it actively shielded the identities and geographical locations of cybercriminals executing data theft and deploying encryption malware⁹². Rashevskyi routinely utilised fabricated identities to lease server space, ignoring abuse complaints to maintain an uninterrupted safe haven for extortionists⁹². Similarly, Silayev developed “cryptors”—sophisticated obfuscation tools designed to disguise malicious payloads as benign files, allowing ransomware to bypass enterprise antivirus detection⁹².

These OFAC sanctions are the culmination of ‘Operation Saffron’, a multi-year, coordinated international law enforcement initiative that saw European authorities and the FBI infiltrate and permanently dismantle the 1VPNS infrastructure in May 2026¹³. By aggressively targeting the service providers, VPN hosts, and tool developers that facilitate cybercrime, allied governments are seeking to systematically dismantle the logistical networks that make industrial-scale ransomware viable¹³.

Emerging Technologies: Bioelectronic Medicine Breakthroughs

While geopolitical conflicts and cyber extortion dominated the enterprise IT landscape, the application of advanced computing within the medical sciences yielded a historic breakthrough. Nature Medicine published the landmark results of a three-year clinical trial conducted by the Feinstein Institutes for Medical Research, detailing the successful implementation of an artificial intelligence-driven “double neural bypass”⁹⁶.

The pioneering procedure was performed on Keith Thomas, a patient left completely paralysed from the chest down following a 2020 spinal cord injury⁹⁸. Surgeons implanted five microelectrode arrays directly into his brain to record neural activity from the motor and sensory cortices⁹⁶. These arrays were integrated with external, non-invasive wearable stimulation patches placed over the muscles of his forearm and spinal cord⁹⁶.

The system relies on highly sophisticated AI algorithms to bridge the severed neurological connection. When the patient formulates the intention to move his hand, the AI decodes the specific neural patterns and instantly translates them into precise electrical stimulation commands delivered to the forearm, effectively moving the paralysed limb⁹⁶. Concurrently, sensors embedded in the hand detect physical contact, routing data back through the AI to stimulate the sensory cortex, artificially recreating the sensation of touch⁹⁸.

The clinical results have been profound. Utilising the technology, the patient regained the ability to feed himself, drink from a cup, and handle delicate objects like empty eggshells without breaking them⁹⁶. Crucially, the intervention demonstrated lasting neuroplasticity; the patient achieved an 86 per cent increase in right arm strength and regained independent tactile sensation in his wrist that persisted even when the system was completely deactivated⁹⁶. This successful integration of machine learning with the human central nervous system represents a monumental leap forward in neuro-restorative therapy, proving that AI can actively facilitate the physiological rewiring of severely damaged neurological pathways⁹⁹.

Conclusions and Strategic Outlook

The global IT industry during the week ending 17 July 2026 illuminated several interconnected, structural shifts that will dictate market dynamics and corporate strategy for the remainder of the decade:

1. The Agentic Compute Bottleneck: The indefinite delay of Google’s Gemini 3.5 Pro starkly exposes the immense difficulty of transitioning from generative text models to autonomous, agentic reasoning. The massive, multi-year capacity expansions announced by TSMC and ASML confirm that the semiconductor supply chain is aggressively retooling to meet the exorbitant compute demands required to overcome this specific software bottleneck.
2. Hardware as the Strategic Moat: The vicious trade-secret litigation between Apple and OpenAI confirms that control over the physical hardware endpoint is paramount. As AI software matures, providers like OpenAI are desperate to bypass legacy smartphone monopolies by building bespoke, AI-native devices. Consequently, incumbent hardware giants will utilise every legal and strategic mechanism available to prevent the exfiltration of their institutional engineering knowledge.
3. The Bifurcation of Digital Governance: The establishment of WAICO in Shanghai solidifies the reality of a fractured global internet. The Global South, wary of the extraterritorial application of Western software embargoes, is actively constructing a parallel digital ecosystem reliant on open-source frameworks and Chinese capacity-building. This ideological divergence will severely complicate international technology compliance and trade.
4. The Industrialisation of Supply Chain Fragility: The catastrophic disruptions at Nichirei Logistics and the exposure of the Kudankulam nuclear blueprints underscore an urgent reality: modern enterprise security postures are fundamentally undermined by the vulnerabilities of third-party vendors. RaaS syndicates like ‘The Gentlemen’ have successfully industrialised the exploitation of this fragility, pivoting toward simultaneous cross-platform encryption to maximise operational paralysis and guarantee extortion payouts.

Moving forward, corporate boards, institutional investors, and sovereign policymakers must cease evaluating technology as an isolated operational expense. Instead, IT infrastructure, foundational models, and cybersecurity must be managed as the primary axes of national security, supply chain resilience, and global geopolitical alignment.

Disclaimer: This report is provided for informational and educational purposes only and does not constitute financial, investment, legal, or regulatory advice. The analysis reflects the data and market conditions available as of the week ending 17 July 2026, and all corporate, market, and geopolitical projections are subject to inherent uncertainties.

Works cited

1. IMF July 2026 WEO: 3.0% Growth and Uneven Tech-Energy Impacts, https://quasa.io/media/imf-july-2026-weo-update-signals-uneven-growth-opportunities-for-investors
2. IMF July 2026 WEO: 3.0% Growth Forecast, https://quasa.io/media/imf-july-2026-weo-update-projects-3-0-global-growth-for-2026
3. TSMC posts record Q2 profit, commits further $100bn to US expansion – Capacity Media, https://capacityglobal.com/news/tsmc-posts-record-profit/
4. What ASML’s Blowout Means for the AI Boom: Nvidia, Micron and TSMC in Focus, https://stocksdownunder.com/asml-ai-boom-nvidia-micron-tsmc/
5. Apple sues OpenAI, saying former Apple employees stole trade secrets – Quartz, https://qz.com/apple-sues-openai-trade-secrets-tang-tan-071026
6. Google’s Gemini 3.5 Pro delayed; Performance falls short of expectations, https://www.businesstoday.in/technology/news/story/googles-gemini-3-5-pro-delayed-performance-falls-short-of-expectations-543524-2026-07-17
7. Mankind creates first global intergovernmental AI cooperation organization, http://en.people.cn/n3/2026/0718/c90000-20479174.html
8. What is WAICO? Inside China’s new 29-nation alliance to counter US AI dominance, https://www.financialexpress.com/business/news/chinas-waico-opens-new-front-in-global-ai-power-race/4295058/
9. AI Legislative Update: July 17, 2026 – Transparency Coalition, https://www.transparencycoalition.ai/news/ai-legislative-update-july17-2026
10. Australian AI Update: Australia changes course | White & Case LLP, https://www.whitecase.com/insight-alert/australian-ai-update-australia-changes-course
11. Nichirei: cyberattack sends shockwaves through Japan’s food supply chain, https://www.salmonbusiness.com/nichirei-cyberattack-sends-shockwaves-through-japans-food-supply-chain/
12. The Gentlemen Ransomware Reaches 478 Victims and Counting – HSM IT Solutions, https://hsmit.com.au/2026/07/10/the-gentlemen-ransomware-reaches-478-victims-and-counting/
13. Sanctioning Ransomware Enablers in Coordinated International Action – State Department, https://www.state.gov/releases/office-of-the-spokesperson/2026/07/sanctioning-ransomware-enablers-in-coordinated-international-action
14. Big Tech Earnings Preview: Is AI Monetization Finally Catching Up to Capex?, https://io-fund.com/ai-stocks/big-tech-earnings-preview-ai-monetization-vs-capex
15. Magnificent Seven Earnings Preview: Can Big Tech Reclaim AI Leadership? – FOREX.com, https://www.forex.com/en-sg/news-and-analysis/magnificent-seven-earnings-preview-can-big-tech-reclaim-ai-leadership/
16. Is ASML Holding (ASML) Expensive Following Its Raised 2026 Revenue Outlook?, https://simplywall.st/stocks/us/semiconductors/nasdaq-asml/asml-holding/news/is-asml-holding-asml-expensive-following-its-raised-2026-rev
17. ASML reports €9.3 billion total net sales and €2.9 billion net income in Q2 2026, https://www.asml.com/en/news/press-releases/2026/q2-2026-financial-results
18. Global Market: ASML raises 2026 outlook as AI boom drives strong earnings, plans capacity expansion, https://m.economictimes.com/markets/us-stocks/news/global-market-asml-raises-2026-outlook-as-ai-boom-drives-strong-earnings-plans-capacity-expansion/articleshow/132412023.cms
19. ASML reports €9.3 billion total net sales and €2.9 billion net income in Q2 2026, https://markets.ft.com/data/announce/detail?dockey=1330-1001212808en-588MO9KVLSK4KDCJJHDLK0UIJJ
20. TSMC vs. NVIDIA: Which AI Semiconductor Stock Should You Buy in July? – TradingView, https://www.tradingview.com/news/zacks:03ee3d1d3094b:0-tsmc-vs-nvidia-which-ai-semiconductor-stock-should-you-buy-in-july/
21. TSMC Raised Its 2026 Revenue Guidance: What It Means for AI Chip Demand – Mitrade, https://www.mitrade.com/au/insights/stock-analysis/asian-stocks/beincrypto-TSM-202607171003
22. New Data Center Developments: July 2026, https://www.datacenterknowledge.com/data-center-construction/new-data-center-developments-july-2026
23. FTI Consulting News Bytes – 17 July 2026, https://fticommunications.com/fti-consulting-news-bytes-17-july-2026/
24. 2026 State of the Digital Decade package | Shaping Europe’s digital future, https://digital-strategy.ec.europa.eu/en/policies/2026-state-digital-decade-package
25. AI News Today July 1 2026: 15 Biggest Stories – Build Fast with AI, https://www.buildfastwithai.com/blogs/ai-news-today-july-1-2026
26. AI explained: Why the world needs to act now | UN News, https://news.un.org/en/story/2026/07/1167848
27. Google Gemini launch delayed as tech falls short of internal goals, Bloomberg News reports, https://live.euronext.com/en/financial-news/google-gemini-launch-delayed-tech-falls-short-internal-goals-bloomberg-news-reports
28. Google Gemini launch delayed as tech falls short of goals – report – TradingView, https://www.tradingview.com/news/forexlive:705208414094b:0-google-gemini-launch-delayed-as-tech-falls-short-of-goals-report/
29. Sudden Halt for Gemini 3.5 Pro: Google’s Flagship AI Faces Troubled Launch Amid Rising User Disappointment – 36氪, https://eu.36kr.com/en/p/3899401765422720
30. Gemini 3.5 Pro Delayed Over Coding, Bloomberg Reports – Search Engine Journal, https://www.searchenginejournal.com/gemini-3-5-pro-delayed-over-coding-bloomberg-reports/582660/
31. Where is Gemini 3.5 Pro? The AI model announced at Google I/O is still MIA. – Mashable, https://mashable.com/tech/google-gemini-3-5-pro-delay
32. Why July 2026’s AI Shake-Up Matters to Sydney Marketing and Tech Teams, https://bigwavedigital.com.au/why-july-2026-s-ai-shake-up-matters-to-sydney-marketing-and-tech-teams/
33. AI News Today July 12 2026: 15 Biggest Stories, https://www.buildfastwithai.com/blogs/ai-news-today-july-12-2026
34. What Apple accuses OpenAI of in explosive trade secret lawsuit? A laptop ‘not returned’, secret interviews, & more, https://www.livemint.com/companies/news/what-apple-accuses-openai-of-in-explosive-trade-secret-lawsuit-a-laptop-not-returned-secret-interviews-and-more-11783762780348.html
35. Apple sues OpenAI, alleging artificial intelligence company stole trade secrets, https://www.theguardian.com/technology/2026/jul/10/apple-sues-openai-trade-secrets
36. What’s behind Apple’s 41-page lawsuit against ChatGPT-maker OpenAI: Trade secrets, talent and the battle for, https://timesofindia.indiatimes.com/technology/tech-news/whats-behind-apples-41-page-lawsuit-against-chatgpt-maker-openai-trade-secrets-talent-and-the-battle-for-/articleshow/132384694.cms
37. Apple Sues OpenAI Over Trade Secret Theft — Amid Musk-Altman Feud, https://bitcoinfoundation.org/news/people-in-crypto-news/apple-vs-openai/
38. Apple sues OpenAI over alleged trade secrets theft – Silicon Republic, https://www.siliconrepublic.com/business/apple-sues-openai-over-alleged-trade-secrets-theft
39. Apple sues OpenAI for allegedly stealing trade secrets | Information Age – ACS, https://ia.acs.org.au/article/2026/apple-sues-openai-for-allegedly-stealing-trade-secrets.html
40. Apple sues OpenAI, alleging trade secret theft for hardware push – Northeast Times, https://northeasttimes.com/2026/07/11/apple-sues-openai-alleging-trade-secret-theft-for-hardware-push/
41. OpenAI and Apple legal battle escalates, poached employees warned about data deletion, https://appleinsider.com/articles/26/07/17/openai-and-apple-legal-battle-escalates-poached-employees-warned-about-data-deletion
42. Report: Apple Sends Legal Letters to Dozens of OpenAI Employees, https://www.macrumors.com/2026/07/17/apple-sends-legal-letters-openai/
43. Apple targets dozens of OpenAI employees with legal letters, https://www.ft.com/content/1b8c9d52-88a9-426b-ba47-f1811f859166?syn-25a6b1a6=1
44. Exclusive-PayPal board sees Stripe-Advent offer as inadequate, sources say, https://www.channelnewsasia.com/business/exclusive-paypal-board-sees-stripe-advent-offer-inadequate-sources-say-6259716
45. Stripe and Advent make $53bn bid for PayPal, https://www.ft.com/content/3738e814-9470-4d7d-94a6-ac5e001a968e?syn-25a6b1a6=1
46. Stripe, Advent offer to buy PayPal for more than $53 billion, sources say, https://www.spokesman.com/stories/2026/jul/14/stripe-advent-offer-to-buy-paypal-for-more-than-53/
47. Stripe and Advent Bid $53 Billion for PayPal Acquisition – Quasa.io, https://quasa.io/media/stripe-and-advent-bid-53-billion-for-paypal-acquisition
48. Stripe and Advent International Make $53 Billion Bid to Acquire PayPal – Stablecoin Insider, https://stablecoininsider.org/stripe-and-advent-international-make-53-billion-bid-to-acquire-paypal/
49. PayPal (PYPL) Stock Forecast: Stripe and Advent Bid $60.50; Is $61.65 the Next Target?, https://www.tradingkey.com/analysis/stocks/us-stocks/262031888-paypal-pypl-stock-forecast-july-15-2026-stripe-advent-60-50-bid-tradingkey
50. Stock Market Today, July 15: PayPal Surges 17% on $60.50 Takeover Bid from Stripe and Advent International | The Motley Fool, https://www.fool.com/coverage/stock-market-today/2026/07/15/stock-market-today-july-15-paypal-surges-17-on-usd60-50-takeover-bid-from-stripe-and-advent-international/
51. Stripe, Advent offer USD 53 bln for PayPal, sources say – The Paypers, https://thepaypers.com/payments/news/stripe-advent-offer-usd-53-bln-for-paypal-sources-say
52. PayPal shares jump over 15% after Stripe, Advent make $53 billion buyout offer, https://m.economictimes.com/markets/us-stocks/news/paypal-shares-jump-over-15-after-stripe-advent-make-53-billion-buyout-offer/articleshow/132418333.cms
53. Stripe and Advent Make $53 Billion Offer for PayPal – CloudColleague, https://cloudcolleague.com/news/53-billion-offer-for-paypal/
54. Above the Fold: Supply Chain Logistics News (July 17, 2026), https://talkinglogistics.com/2026/07/17/above-the-fold-supply-chain-logistics-news-july-17-2026/
55. Indonesia joins China-initiated global AI cooperation body, https://en.antaranews.com/news/422999/indonesia-joins-china-initiated-global-ai-cooperation-body
56. 29 countries sign agreement on establishing World AI Cooperation Organization, https://english.www.gov.cn/news/202607/17/content_WS6a59a226c6d00ca5f9a0c432.html
57. Guterres: AI must be shaped by ‘all of humanity’, not a handful of powers, https://news.un.org/en/story/2026/07/1167965
58. Chair’s Statement of the 2026 World Artificial Intelligence Conference & High-Level Meeting on Global AI Governance_Ministry of Foreign Affairs of the People’s Republic of China, https://www.mfa.gov.cn/eng/xw/zyxw/202607/t20260717_11984715.html
59. China’s Xi calls for step up of global effort in AI, as US curbs squeeze China’s tech access, https://www.klcc.org/npr-news/2026-07-17/chinas-xi-calls-for-step-up-of-global-effort-in-ai-as-us-curbs-squeeze-chinas-tech-access
60. President Xi Jinping Attends the Opening Ceremony of the 2026 World AI Conference and High-Level Meeting on Global AI Governance and Delivers Keynote Address_Ministry of Foreign Affairs of the People’s Republic of China, https://www.mfa.gov.cn/eng/xw/zyjh/202607/t20260717_11984910.html
61. President Xi Jinping Attends the Opening Ceremony of the 2026 World AI Conference and High-Level Meeting on Global AI Governance and Delivers Keynote Address, https://sydney.china-consulate.gov.cn/eng/zgxw/202607/t20260717_11984910.htm
62. Navigating AI bubble: Balancing innovation and policy in Indonesia, https://en.antaranews.com/news/422596/navigating-ai-bubble-balancing-innovation-and-policy-in-indonesia
63. From Principles to Power Points: What the Government’s “AI in Australia’s Interests” means for business – Clayton Utz, https://www.claytonutz.com/insights/2026/july/from-principles-to-power-points-what-the-governments-ai-in-australias-interests-means-for-business
64. On our terms: Australia announces world-first AI framework, https://minterellison.co.nz/insights/on-our-terms-australia-announces-world-first-ai-framework
65. Office of AI | PM&C – Department of the Prime Minister and Cabinet, https://www.pmc.gov.au/domestic-policy/office-ai
66. AI (artificial intelligence) | The Guardian, https://www.theguardian.com/technology/artificialintelligenceai
67. Big Tech promises the world with AI. Australians shouldn’t believe them., https://australiainstitute.org.au/post/big-tech-promises-the-world-with-ai-australians-shouldnt-believe-them/
68. Federal budget 2026: The biggest tech announcements | Information Age – ACS, https://ia.acs.org.au/article/2026/federal-budget-2026-the-biggest-tech-announcements.html
69. Prepare for the AusAlert national test on 27 July 2026 – NEMA, https://www.nema.gov.au/about-us/media-centre/prepare-ausalert-national-test-27-july-2026
70. Government Agencies Falling Victim to Ransomware Daily, Warns Study, https://www.infosecurity-magazine.com/news/government-ransomware-daily/
71. Compromised Logins Surge as the Most Common Entry Point for Ransomware Attacks, https://www.infosecurity-magazine.com/news/compromised-logins-ransomware-entry/
72. Cyber Brief: NCSC warns on router targeting, identity attacks rise – Secarma: Penetration Testing and Cybersecurity Company, https://secarma.com/16-07-2026-cybersecurity-news-roundup
73. No Manners Here: The Ruthless Rise of The Gentlemen Ransomware, https://unit42.paloaltonetworks.com/the-gentlemen-ransomware/
74. The Gentlemen ransomware: Dissecting a self-propagating Go encryptor | Microsoft Security Blog, https://www.microsoft.com/en-us/security/blog/2026/05/28/the-gentlemen-ransomware-dissecting-a-self-propagating-go-encryptor/
75. The Gentlemen Overtakes Qilin as Most Prolific Ransomware Threat, https://www.infosecurity-magazine.com/news/the-gentlemen-most-prolific/
76. The Gentlemen ransomware targets Windows, Linux, ESXi and NAS – Paubox, https://www.paubox.com/blog/the-gentlemen-ransomware-targets-windows-linux-esxi-and-nas
77. The Gentlemen: When Ransomware-as-a-Service Becomes an Industry, https://incyber.org/en/article/the-gentlemen-when-ransomware-as-a-service-becomes-an-industry/
78. A cyberattack hit Nichirei, one of Japan’s largest food companies, https://securityaffairs.com/195543/security/a-cyberattack-hit-nichirei-one-of-japans-largest-food-companies.html
79. Nichirei getting back online after cyberattack hit KFC supplies – The Japan Times, https://www.japantimes.co.jp/business/2026/07/16/companies/nichirei-back-cyberattack/
80. Nichirei cyberattack disrupts food and cold chain operations as Kudankulam data leak flags rising infrastructure threats, https://industrialcyber.co/critical-infrastructure/nichirei-cyberattack-disrupts-food-and-cold-chain-operations-as-kudankulam-data-leak-flags-rising-infrastructure-threats/
81. Nichirei to Fully Resume Operations as Early as Next Week, https://www.nippon.com/en/news/yjj2026071700736/
82. Cyberattack on Japan’s largest cold-chain operator disrupts KFC, supermarket supplies, https://therecord.media/cyberattack-japan-nichirei-logistics-impacts-kfc
83. Cyberattack hits Japan’s ice cream giant Glico after KFC, disrupts sushi and supermarket deliveries, https://timesofindia.indiatimes.com/business/international-business/cyberattack-hits-japans-ice-cream-giant-glico-after-kfc-disrupts-sushi-and-supermarket-deliveries/articleshow/132441658.cms
84. Preliminary Investigation Report: Data Extortion targeting Kudankulam Nuclear Plant engineering documents – Shieldworkz, https://shieldworkz.com/blogs/preliminary-investigation-report-data-extortion-targeting-kudankulam-nuclear-plant-engineering-documents
85. Exclusive-Files relating to India’s largest nuclear power plant Kudankulam exposed in data breach | WTVB | 1590 AM, https://wtvbam.com/2026/07/15/exclusive-files-relating-to-indias-largest-nuclear-power-plant-kudankulam-exposed-in-data-breach/
86. ‘Serious risk’ to India’s largest nuclear plant after sensitive files leaked on dark web: Report, https://timesofindia.indiatimes.com/india/serious-risk-to-indias-largest-nuclear-plant-after-sensitive-files-leaked-on-dark-web-report/articleshow/132413344.cms
87. Kudankulam nuclear plant files on the dark web: Inside the alleged data leak, https://indianexpress.com/article/explained/explained-economics/kudankulam-files-on-the-dark-web-inside-the-alleged-data-leak-10790059/
88. Kudankulam nuclear plant data breached, NPCIL says core systems untouched, https://timesofindia.indiatimes.com/india/kudankulam-nuclear-plant-data-breached-npcil-says-core-systems-untouched/articleshow/132425150.cms
89. Ransomware attack forces Coca-Cola to suspend US production at dairy unit, https://www.cybersecuritydive.com/news/ransomware-attack-coca-cola-suspend-production-dairy/825540/
90. Who’s Hacked? Latest Data Breaches And Cyberattacks – Cybercrime Magazine, https://cybersecurityventures.com/intrusion-daily-cyber-threat-alert/
91. EY Data Breach – Hackers Gain Access to IT Support System and Download Documents, https://cybersecuritynews.com/ey-data-breach/
92. Treasury Sanctions Malware and Infrastructure Providers Supporting Ransomware Attacks Against Americans, https://home.treasury.gov/news/press-releases/sb0559
93. Treasury sanctions VPN provider, individuals tied to hospital ransomware, https://www.beckershospitalreview.com/healthcare-information-technology/cybersecurity/treasury-sanctions-vpn-provider-individuals-tied-to-hospital-ransomware/
94. Could Your Network Be Next? What the New Ransomware Sanctions on 1VPNS and Crypter Sellers Mean for You – Cloaked, https://www.cloaked.com/post/could-your-network-be-next-what-the-new-ransomware-sanctions-on-1vpns-and-crypter-sellers-mean-for-you
95. US Sanctions First VPN in Crackdown on Ransomware Criminals – CNET, https://www.cnet.com/tech/services-and-software/us-treasury-sanctions-first-vpn-service-1vpns-ransomware/
96. Feinstein Institutes’ double neural bypass trial featured in Nature Medicine, https://feinstein.northwell.edu/news/the-latest/double-neural-bypass-technology-featured-on-cover-of-nature-medicine
97. Feinstein Institutes Technology That Restores Movement and Sensation After Paralysis Featured in Nature Medicine, https://lasvegassun.com/news/2026/jul/16/feinstein-institutes-technology-that-restores-move/
98. Brain implant helps paralysed man to feed himself and drink from cup, https://www.theguardian.com/science/2026/jul/16/neural-bypass-brain-implant-paralysed-man-feed-himself-drink-from-cup
99. Double neural bypass: Restoring movement, touch in paralysis patients | Northwell Health, https://feinstein.northwell.edu/news/insights/developing-double-neural-bypass-restore-lasting-movement-sensation-paralysis
100. Paralyzed Man Feels Sensation Again With Brain Stimulation Device | The Scientist, https://www.the-scientist.com/paralyzed-man-feels-sensation-again-with-brain-stimulation-device-74754
101. AI and a brain implant restored a paralysed man’s movement and touch – TNW, https://thenextweb.com/news/feinstein-double-neural-bypass-bci-paralysis-nature-medicine