OpenAI, Google, and Anthropic are setting the pace in artificial intelligence at the start of December 2025. The atmosphere in Silicon Valley and Beijing is tense. OpenAI has declared a “Code Red” in response to mounting competitive pressure, while new models are emerging: Claude Opus 4.5, Gemini 3, and DeepSeek-V3.2. Simultaneously, airlines, banks, and retail chains are signing record-breaking AI deployment deals. Yet, rapid progress hides a serious problem—79% of companies fall short of global AI security standards, according to a December report from Cambridge researchers. What’s fueling this technological race, and why is the pace so relentless?
New AI Models Are Changing the Game
Anthropic unveiled Claude Opus 4.5—a model that outperformed rivals in programming and business process automation tests. Opus 4.5 introduces new defenses against prompt injection attacks and tools for agent developers. The result? Claude Code, Anthropic’s programming agent, generated a billion dollars in revenue in just six months. The company also acquired Bun to strengthen its codebase. Google isn’t falling behind. Gemini 3 surpassed 1501 Elo on LMArena and scored 91.9% in the GPQA Diamond test, placing it at the top of the rankings. Gemini 3 Pro has been integrated into Search, Workspace, and developer tools, while the Antigravity platform aims to accelerate agentic coding. Under pressure, OpenAI is delaying new feature launches, focusing on evolving ChatGPT and releasing GPT-5.1 with adaptive “thinking” time and 24-hour prompt caching. Behind the scenes, the Garlic model is in development—a direct response to Google’s offensive.
Business Sector Focuses on Partnerships and Infrastructure
Chinese and open-source models refuse to play second fiddle. DeepSeek-V3.2, with 685 billion parameters, matches GPT-5 and Gemini-3.0-Pro in math and programming, and its sparse attention architecture has slashed the cost of processing 128,000 tokens by 70%. Meanwhile, Kimi K2 Thinking from Moonshot AI, equipped with a trillion parameters, can execute up to 300 tool operations in a single reasoning chain.
New alliances are emerging. Virgin Australia and Wesfarmers are deploying OpenAI for customer service and logistics, while HSBC has signed a multi-year deal with French player Mistral, boosting Europe’s AI standing. Google is challenging Nvidia—analysts predict that by 2030, Google’s TPUs will capture 25% of the AI chip market, which is expected to exceed $440 billion. Google shares are up 66%. Meta is negotiating to buy Google’s chips for 2026–2027—a move towards infrastructure diversification.
AI Market: Investments, Restructuring, and the Chip Race
In manufacturing, Neurologik debuts AI agents for precision process management, blending product logic, safety standards, and historical data. Microsoft is betting on Work IQ, an analytics layer enabling Copilot to recognize user work patterns and company knowledge in documents and emails.
Security and Regulation: The Industry at a Crossroads
November and December 2025 saw over $3.5 billion invested in AI—a record set by Mistral AI (€2 billion). TandemAI has raised $22 million for drug discovery algorithms, Sweet Security secured $75 million for its security platform, and WisdomAI landed $50 million for next-gen analytics. Meanwhile, the industry is already spending $88 billion on data centers and compute power. OpenAI and AWS are investing $38 billion, while Microsoft and NVIDIA are putting $15 billion into Anthropic’s growth. HP has announced a restructuring: by 2028, 4,000–6,000 jobs will be cut, with resources redirected toward developing AI-PCs.
Security remains a critical issue. The December report reveals most companies aren’t meeting global standards. For the first time, a cyberattack was executed by an autonomous AI system—an event that sent shockwaves through the industry. Uzbekistan has launched a national AI strategy, investing in a Nvidia supercomputer and opening a transfer office in Silicon Valley. Australia is adopting a strategy based on existing regulations, investing in data centers and training. Europe is debating new regulations, while EY and NVIDIA are helping companies test and implement physical AI systems in line with compliance requirements.
Agentic AI and European Independence: Trends for 2026
Agentic AI is taking center stage. Google Antigravity and Microsoft Agent 365 demonstrate that the market is shifting from passive tools to autonomous digital workers. Amazon, Google, and OpenAI are racing for dominance in AI-powered online commerce—shopping via chatbot could upend the current e-commerce model. Europe is strengthening its position—Mistral and its partnership with HSBC show the region isn’t ready to cede ground to the US and China.
Who will gain the upper hand in the coming weeks? OpenAI is playing defense, Google is merging models and chips into a single offering, and Anthropic is betting on security and developer tools. The AI sector is undergoing its biggest transformation since the launch of ChatGPT—and we’ll know the winners sooner than you think.