ET highlights 25 pivotal advancements in a transformative year for artificial intelligence (AI) while casting an eye toward the forthcoming year.
Investments, Circular Funding Dynamics, Bubble Concerns
1. OpenAI reaches $500 billion valuation, transitions to a for-profit model, and separates from Microsoft
OpenAI achieved a groundbreaking $500 billion valuation, becoming the largest privately-held enterprise in the AI sector, fueled by significant fundraising initiatives. Transitioning to a for-profit status, the relationship with Microsoft has reportedly grown tenuous, as OpenAI seeks to assert its autonomy.
This shift is reshaping the AI landscape, enabling both entities to explore expansive partnerships. Notably, OpenAI is reportedly eyeing a $1 trillion IPO, one anticipated to generate immense attention in 2026.
2. Circular funding raises bubble apprehensions; Nvidia’s market cap declines
A circular funding ecosystem has emerged within AI, with Nvidia, Microsoft, Oracle, AMD, SoftBank, and OpenAI investing reciprocally. Over eight years, OpenAI has engaged in transactions amounting to approximately $1.4 trillion.
Wall Street has begun to highlight concerns surrounding inflated demand for data centers and debt-laden agreements. As a result, fears of an AI bubble have intensified, particularly following Nvidia’s stock market tremors, despite it being the first corporation to amass a market cap of $5 trillion.
3. AI startups capture over half of global VC funding
In 2025, AI startups commanded an unprecedented 53% of the $400 billion global venture capital, marking the highest proportion recorded. Historically, enterprise startups aspired to achieve $1 million in annual recurring revenue (ARR) in their inaugural year.
Currently, an average AI company surpasses this threshold with $2.1 million, while the elite 1% exceed $5.3 million. Remarkably, these firms are securing Series A funding a mere seven months post-monetization—a pace unparalleled in startup history.
4. AI co-founders emerge as the world’s youngest billionaires
The trio behind AI startup Mercor—Adarsh Hiremath, Brendan Foody, and Surya Midha—has made headlines as the youngest self-made billionaires at the age of 22, overshadowing Mark Zuckerberg’s longstanding record.
All three are college dropouts and Thiel Fellows who have cultivated a talent-sourcing platform, skyrocketing its valuation to $10 billion within a mere two years.
5. Big Tech intensifies M&A for AI supremacy
Leading technology firms have escalated consolidation efforts. Google acquired coding platform Windsurf, while Meta pivoted towards a “superintelligence” initiative, investing $14.3 billion for a 49% stake in Scale AI.
Scale’s founder, Alexandr Wang, now spearheads Meta’s Superintelligence Labs, while Nvidia has acquired chip inference startup Groq for $20 billion.
2026 Projections
The AI realm is poised to transition from hyperbole to tangible economic realities, prompting a valuation recalibration in both private and public arenas.
It is anticipated that startup valuations will stabilize alongside a surge in M&A activity, with established tech giants seeking to acquire startups for talent, data, and defendable intellectual property.
OpenAI and Anthropic may both attain a valuation of $1 trillion and are expected to contemplate IPOs in the coming year. A pivotal consideration remains whether enterprise demand can sustain the influx of capital directed toward data centers.
AI Model Development, China, and DeepSeek
6. The Dawn of DeepSeek
The introduction of DeepSeek in 2025 signified a notable advancement. This economical Chinese model employed reinforcement learning, a machine learning methodology, substantially slashing training costs. Its arrival challenged the prevailing belief that extensive capital is requisite for large language model (LLM) training.
7. Vibe Coding and Browser Rivalries
Yearly themes revolved around reasoning and agentic AI capabilities, with vibe coding crystallizing as a notable trend. A plethora of new models aimed to enhance coding functionalities. The focus also pivoted toward AI-driven commerce, witnessing innovations like OpenAI’s Atlas and Google’s Gemini Shopping.
8. Chinese AI models achieve parity
Chinese open-source AI models, exemplified by Moonshot AI’s Kimi and Alibaba’s Qwen, have demonstrated performance levels comparable to their American counterparts, notwithstanding US restrictions on the export of high-end Nvidia chips.
9. Viral Phenomena: Ghibli and Nano Bananas
Visual content generated by Ghibli and Nano Bananas has spurred numerous viral sensations. Tools such as Nano Banana and Sora have democratized image editing for non-technical users, sparking debates surrounding copyright and privacy amid a regulatory vacuum.
10. The Rise of Superintelligence
Superintelligence has gained traction, with prognostications indicating that proximity to artificial general intelligence (AGI) has markedly increased.
Corporations like Microsoft and Meta have established specialized teams to delve into various facets of superintelligence, while OpenAI recently asserted a commitment to realize such a breakthrough within a decade, despite growing calls to halt development.
2026 Outlook
James Landay, co-director of Stanford HAI, forecasts that AGI is improbable by 2026; however, sovereign AI models are likely to garner significant focus. An increase in exploration of alternatives to transformer-led models, foundational for LLMs like ChatGPT, is anticipated.
Yann LeCun, formerly of Meta AI, has departed to launch Advanced Machine Intelligence, targeting world AI models. Google and Fei-Fei Li, recognized as the AI matriarch and co-founder of World Labs, are pursuing similar objectives.
Jobs, Layoffs, Talent Competition, and Work Future
11. Unprecedented job losses since the Internet boom
In the wake of AI-driven restructuring and a profit-oriented push, 2025 emerged as the most significant year for tech layoffs since the internet surge. Over 122,000 job eliminations were documented across 257 companies, according to the independent monitoring platform Layoffs. fyi.
Amazon dismissed 14,000 employees, Intel cut its workforce by 24,000, and Microsoft let go of 9,000. Accenture, Oracle, Meta, and Salesforce also downsized teams, attributing decisions to AI-facilitated productivity enhancements.
12. Lucrative compensation packages from Meta, OpenAI, etc.
The scramble for talent in Silicon Valley intensified, with Meta and OpenAI offering extravagant financial packages, including signing bonuses soaring to $100 million for elite researchers, even as layoffs persisted throughout the broader tech arena.
13. TCS faces layoffs for the first time in history
Tata Consultancy Services (TCS), a titan in India’s IT sector, laid off approximately 2% of its workforce (12,000 employees) in July 2025, amid global market uncertainties and the needs of an AI-centric organization. Numerous other firms similarly reported quiet layoffs as automation took center stage.
14. Widespread AI-driven employee screening
AI has emerged as a double-edged sword in recruitment. Globally, recruiters are adopting AI-based tools, with around 75% of Indian companies allocating up to 70% of their hiring budgets to AI-infused recruitment methodologies, per a LinkedIn survey.
India’s leading IT firms have commenced AI-driven screening for approximately 70-80% of applications while utilizing chatbot-led interviews during preliminary evaluations.
15. The era of the hybrid workforce
Marc Benioff, CEO of Salesforce, remarked that today’s leaders constitute the “last generation of executives directing solely human workforces,” signaling the dawn of a hybrid workforce—with human employees operating alongside digital agents.
2026 Projections
Hiring trends among IT firms are anticipated to remain cautious, with many traditional roles devoid of AI/ML likely to stagnate, as per recruiting firms.
The World Economic Forum Future of Jobs Report 2025 predicts a transformation across 23% of jobs by 2027, driven by AI and automation, a phenomenon expected to unfold throughout 2026.
AI Infrastructure, Compute Conflicts, Hyperscaler Capital Expenditure
16. Hyperscalers allocate over $400 billion in annual capital expenditure for AI infrastructure
In 2025, hyperscale firms invested an unprecedented $400 billion in capital expenditures, marking the most extensive allocation to date.
Tech giants, including Microsoft, Amazon, Google, and Meta, are developing expansive data centers equipped with advanced GPUs to accommodate the growing demand for generative AI services and model training.
17. Environmental implications of infrastructure expansion
The proliferation of AI infrastructure presents significant environmental challenges. The International Energy Agency (IEA) reported in 2025 that global data center electricity consumption is predicted to more than double from 460 terawatt-hours in 2022 to 945 terawatt-hours by 2030, with AI responsible for 35% to 50% of this consumption.
18. Google’s aspiration for space-based data centers
Google unveiled Project Suncatcher with ambitions to explore space-based data center technologies. The initiative is set to launch two solar-powered satellites in early 2027, each housing four TPUs capable of conducting AI model training leveraging solar energy.
19. The battle for custom AI chips heightens
The competition in AI chip development has intensified as hyperscalers accelerate their endeavors to create bespoke silicon. Products include Google’s TPUs, Amazon’s Trainium and Inferentia, Microsoft’s Maia, and Huawei Ascend. Reports of Meta’s agreement to utilize Google’s TPUs have cast clouds over Nvidia’s market position.
20. Decline in token expenses
The cost of tokens plummeted by nearly 90%, from $20 per million tokens in 2022 to a mere $0.40 in 2025, driven by enhanced hardware efficiency and software optimization, in line with Moore’s Law.
OpenAI’s CEO, Sam Altman, posits that AI costs are decreasing tenfold annually. This phenomenon has rejuvenated interest in Jevons’ Paradox, as more affordable technology encourages a surge in usage.
2026 Outlook
Capital expenditure by leading hyperecalers—Google, Microsoft, Meta, Amazon, Oracle—is forecast to surpass $600 billion in 2026, a 36% escalation from 2025. Approximately 75% of this expenditure, roughly $450 billion, will be directly associated with AI infrastructure rather than traditional cloud services.
Google’s TPUs could be primed to erode Nvidia’s supremacy, with projected shipments anticipated to reach 2.7 million units by 2026 compared to Nvidia’s 9.4 million. Concerns surrounding sustainability will prompt a reliance on nuclear energy, solar power, and satellites to fuel tomorrow’s data centers.
India’s AI Narrative
21. India inaugurates Rs10,000 crore AI Mission; compute costs dip below $1
With a budget exceeding Rs 10,000 crore, the Indian administration has launched the IndiaAI Mission, designed to augment AI infrastructure.
In contrast to the international pricing benchmark of $2.5 to $3 per hour, the initiative has successfully deployed over 38,000 GPUs, surpassing its initial goal of 10,000. India boasts the most economical AI compute, available for less than Rs 70 per hour, as a result of a 40% government subsidy.
22. Global AI titans establish operations in India
International AI stakeholders are laying down roots in India’s burgeoning landscape. OpenAI is setting up an office in New Delhi, Anthropic is planning its inaugural Indian office in Bengaluru, and Figma has also commenced operations there, along with others like ElevenLabs announcing similar expansions. With over 900 million internet users and 90% engaging with AI-driven applications, India has emerged as the largest user base for myriad AI firms.
23. Major investments in data centers from Reliance, Microsoft, Amazon, Google, Adani, and TCS
Significant investments have been committed to data centers by global hyperscalers and leading Indian conglomerates, projecting over $50 billion in expenditure in the next five to seven years, elevating India’s total capacity to approximately 9-10 GW from the current 1 GW.
24. Developer community in India reaches 22 billion; expected to surpass the US by 2028
With 21.9 million professionals, India now boasts the second-largest developer community globally on GitHub. Projections from GitHub suggest that India will eclipse the US, attaining more than 57.5 million developers by 2030.
25. Indic models: Sarvam, Gnani, et al
As part of the IndiaAI Mission, Indian startups like Sarvam AI and Gnani AI are pioneering foundational AI models for Indic languages. Enterprises like TechM are also developing Indic language models as foreign innovations enhance their capabilities in this domain.
2026 Outlook
The trajectory of India’s AI ecosystem is projected to progress from a testing phase to substantive implementation in 2026. Anticipations surrounding sovereign AI are poised to intensify, alongside expectations for an effective policy framework.
The nation has already initiated an AI framework and is reportedly deliberating an AI law. The upcoming India Impact Summit in February 2026 is set to attract leading figures in AI to New Delhi, generating considerable global interest.
Source link: M.economictimes.com.
