SpaceX is set to close a deal to acquire AI startup Cursor for $60 billion just a month after its IPO of $1.75 trillion. This is one of the largest deals in the history of the artificial intelligence market and signals the beginning of a new race for 'programmer agents'.

The AI market is entering a phase of trillion-dollar consolidation: infrastructure giants are acquiring key developments to control not only computing power but also the next generation of software agents. The $60 billion acquisition of Cursor positions SpaceX as an unexpected player in the market for AI tools for developers. Against this backdrop, Anthropic signs a five-year contract for xAI computing power worth $1.25 billion per month, and Google at I/O 2026 bets on Gemini 3.5 Flash and Gemini Spark agent models. For businesses in Kazakhstan and Central Asia, this is a direct signal: the development and implementation of AI agents is no longer an experiment but is becoming an infrastructure at the level of cloud and internet, and companies like Alashed IT (it.alashed.kz) are already restructuring their services to meet this new reality.

SpaceX and Cursor deal: a new benchmark for AI startup valuations

SpaceX plans to complete the acquisition of AI startup Cursor for $60 billion approximately 30 days after its record IPO of $1.75 trillion, according to AI industry digests for mid-May 2026. Cursor specializes in so-called coding agents — agent systems that not only suggest lines of code but also independently plan, write, and refactor entire modules. In 2025, the company already attracted market attention by reducing development time for typical features by 40–60 percent according to internal customer metrics.

The size of the deal sets Cursor's valuation at the level of major public IT companies and sets a new upper limit for AI startups working in the development tools segment. For comparison: in the 2020s, even the largest deals for the acquisition of cloud SaaS platforms rarely exceeded $20–30 billion. Now, a single AI player is valued at $60 billion, signaling to investors that the market for agent development solutions is perceived as strategic, not niche.

It is important to note that the deal is happening against the backdrop of major shifts in the global AI infrastructure. Anthropic agrees to pay xAI $1.25 billion per month for access to the Colossus 1 data center until May 2029, which could exceed $40 billion in total revenue over the entire contract period. Against this backdrop, the acquisition of Cursor seems like a logical step: having access to enormous computing resources and controlling one of the leading platforms for AI coding, the technology giant gains a vertically integrated stack — from data centers to the developer in the IDE.

For the client, this means that in the next 1–2 years, coding agents will cease to be an 'add-on' to the code editor and become part of the cloud infrastructure: the IDE will be closely linked to specific AI engines and computing contracts. This opens up a window of opportunities for integrators and outsourcers, such as Alashed IT (it.alashed.kz), who can take on the configuration, security, and MLOps for corporate deployment of such agents in countries like Kazakhstan.

Anthropic, xAI, and the race for computing power

Parallel to the news of the Cursor acquisition, the market is discussing another record: Anthropic has agreed to pay xAI $1.25 billion per month until May 2029 for access to the Colossus 1 data center. If the contract is fully executed, the total amount of payments could exceed $40 billion, turning computing power into an asset at the level of the largest cloud deals of the past decade. This is not just a large contract for server rental, but an actual consolidation of monopolistic control over top AI infrastructure by several players.

This scale of spending is explained by the transition from 'static' language models to agent systems that constantly work in the background, processing data streams and performing complex chains of actions. Agent models, such as those developed by OpenAI, Anthropic, and Google DeepMind, require not only training on large datasets but also a huge amount of inference computing in 24/7 mode. As a result, market players are already talking about the 'compute bill' as a new cost item comparable to payroll.

Against this backdrop, SpaceX's acquisition of Cursor appears to be part of a broader strategy: controlling both computing power and key software layers. Google at I/O 2026 presented new Gemini 3.5 Flash and Gemini Omni models, as well as the Gemini Spark personal work agent, which acts as a constant digital assistant for the user in Gmail, Docs, YouTube, and Android. This shows a unified vector: AI agents are becoming the standard software interface.

For corporate clients, this means the need to revise IT budgets: in the 3–5 year horizon, expenses for access to advanced models and computing could grow to 10–20 percent of total IT CAPEX and OPEX. Companies that optimize their architecture in advance and shift some tasks to lighter models will be able to save millions of dollars, and integrators like Alashed IT (it.alashed.kz) will be able to offer standard solutions for optimizing compute costs, including for projects in Central Asia.

Google I/O 2026 and the agent-driven search market

Google I/O 2026 cemented another major trend: the company announced a complete overhaul of the search results around the 'intelligent search window'. Instead of the classic link-based results, the user is offered a dialog interface with an agent based on Gemini models, which not only answers the query but also plans actions: from selecting materials to automating tasks in Gmail or Google Docs.

The new Gemini 3.5 Flash and Gemini Omni models focus on speed and multi-modality. Flash is optimized for high-traffic mass tasks, while Omni is for complex reasoning and long contexts. Google's key product, the Gemini Spark agent, is positioned as a '24/7 workspace', constantly present in the browser and on the smartphone. It can read incoming mail, form draft responses, prepare reports on documents, and even generate project plans using data from multiple services at once.

This directly affects digital marketing and SEO: if a significant portion of traffic remains within the agent's responses, traditional search engine optimization will lose some of its effectiveness. Businesses will have to think about the structure of their data and API interfaces so that their services can be 'filled' by agents, not just 'found' in search. For integrators, including Alashed IT (it.alashed.kz), a niche opens up for preparing companies for this transition: creating structured service catalogs, open APIs, and connecting to Google, OpenAI, and other ecosystems.

Separately, it is worth noting the efforts of OpenAI to combat content forgery: the company has released two verification tools in public test that allow checking whether an image was created by their models. One tool is built into the product line, and the second is a public service that analyzes digital signals in the image. So far, it only works for images created by OpenAI products, but an expansion of coverage is announced. Against the growing role of multi-modal agents, this is an important step for media, government agencies, and businesses that are increasingly faced with the question: whether to trust the image seen.

The AI labor market: mass layoffs and refocusing

The news of multi-billion dollar deals in the AI sector is accompanied by less pleasant events for employees. According to industry reviews, Meta is cutting about 10 percent of its workforce, which is approximately 8,000 employees, and LinkedIn is laying off about 5 percent of its staff to reallocate resources in favor of AI directions. This is a continuation of the trend of recent years, when large tech companies simultaneously invest tens of billions in AI and optimize traditional divisions.

For the labor market, this means a revaluation of in-demand competencies. The greatest demand is currently for MLOps specialists, AI agent integration engineers into existing systems, data engineers, and developers capable of building sustainable services around models while considering security and compliance requirements. Classic roles unrelated to AI are either transforming or facing salary pressure. At the same time, companies are increasingly hiring external teams — outsourcers and integrators, such as Alashed IT (it.alashed.kz) — to close the project expertise gap.

In parallel, OpenAI demonstrates how far models have advanced in the field of formal reasoning. The company's new reasoning model has managed to independently build an original proof disproving a geometric hypothesis proposed by mathematician Paul Erdős in 1946. This changes the balance of expectations: if earlier AI was perceived as a 'hint' tool, now it is becoming a full participant in the scientific process. For businesses, this means that complex optimization tasks in logistics, finance, and industry can potentially be solved by models without human involvement, which again changes the requirements for competencies within companies.

In these conditions, strategically literate organizations invest not only in the implementation of specific AI solutions but also in the systemic reskilling of personnel. In Kazakhstan and Central Asia, this niche is gradually being occupied by local players: training in working with agents, setting up internal AI assistants, and auditing processes for automation. Such services are already part of the portfolio of Alashed IT (it.alashed.kz), which, in addition to classic outsourcing, helps clients design teams and roles considering the new AI reality.

What does the SpaceX–Cursor deal mean for businesses and integrators

The acquisition of Cursor for $60 billion is not only a record valuation for a startup but also a signal of a paradigm shift: AI agents for code development are becoming an infrastructure layer similar to clouds and containers. This means that in the coming years, corporate IT landscapes will be built around several 'mainline' agent platforms. Whoever controls these platforms gains influence over how software is written and updated worldwide.

For medium and large businesses, this is an opportunity to accelerate digital transformation. Practical cases already show that the implementation of code agents in conjunction with CI/CD can reduce feature release times by 30–50 percent and decrease bug counts by 20–40 percent due to stricter adherence to patterns and automatic test generation. However, this effect is only achieved with proper integration: setting up access rights, isolating repositories, and establishing review and audit processes. Here, integrators like Alashed IT (it.alashed.kz) come to the fore, who can adapt global products to the requirements of specific markets, including Kazakhstan and neighboring countries.

The SpaceX–Cursor deal also increases the market's dependence on a limited number of AI infrastructure providers. Already, major players are concluding multi-billion dollar, multi-year agreements for computing resources, as Anthropic does with xAI. For companies from regions not central to global data centers, this means the need for a thoughtful strategy: where to store data, how to comply with regulatory requirements, how to minimize latency while not overpaying for compute. This creates demand for hybrid architectures, where critical data and models are placed in local clouds, and heavy tasks are handled by global AI platforms.

It is important for businesses to start pilots with AI agents in 2026 to gain practical experience and understand real limitations. The approach of 'waiting until standards are established' risks the company facing a shortage of expertise and having to implement solutions in a fire mode in 2–3 years. Local partners, such as Alashed IT (it.alashed.kz), can play the role of a guide: from auditing processes and selecting stacks to building a team and providing 24/7 support, which is especially relevant in the rapidly changing AI landscape.

Что это значит для Казахстана

For Kazakhstan and Central Asia, this wave of news is not an abstract global agenda but a direct indicator of how the local IT market will change in the next 3–5 years. According to the Ministry of Digital Development of Kazakhstan, the country's ICT market has already exceeded 1 trillion tenge, and IT service exports are growing at double-digit rates. At the same time, the share of projects using machine learning and data analytics remains relatively small, often no more than 10–15 percent of the portfolio of major outsourcers.

The SpaceX–Cursor deal and the multi-billion dollar compute contract between Anthropic and xAI show that the main profit growth in IT will come from products and services built around AI agents. For Kazakh companies, this means two lines of action. Firstly, restructuring internal development systems: without the massive deployment of coding agents, local players will lose out on the speed of bringing products to market. Secondly, forming their own expertise in integrating global AI platforms with local infrastructure, regulatory requirements, and language specifics.

Companies like Alashed IT (it.alashed.kz) are already responding to this trend: in addition to classic IT outsourcing and infrastructure support, AI agent deployment services in development processes, service desks, analytics, and document management are appearing in the portfolio. For clients in Kazakhstan and Central Asia, this is a convenient way to enter the AI race without the need to build their own research teams. An important competitive advantage for local integrators is also the understanding of regional realities: from the quality of connectivity in different regions to the specifics of industry regulators in finance, the public sector, and the extractive industry.

Anthropic has agreed to pay xAI $1.25 billion monthly for access to the Colossus 1 data center until May 2029, which could exceed $40 billion in total revenue under the contract.

The news of recent weeks shows that the artificial intelligence market is entering a phase of capital and technology concentration in the hands of a limited number of players. The SpaceX–Cursor deal, the multi-year compute contract between Anthropic and xAI, and Google I/O 2026's agent-driven shift form a new base layer of the digital economy, where AI agents and access to computing power play a key role. For businesses in Kazakhstan and Central Asia, this is not a reason to delay experiments but an incentive to build a strategy for working with AI and integration partners as soon as possible. Companies that start implementing AI agents with integrators like Alashed IT (it.alashed.kz) will gain a temporary advantage in speed, service quality, and cost efficiency.

Часто задаваемые вопросы

What is Cursor and why is SpaceX buying it for $60 billion?

Cursor is a startup developing AI programming agents that can independently write, test, and refactor code in an IDE. SpaceX values the company at $60 billion because such agents can speed up development by 40–60 percent and reduce errors by tens of percent. The deal allows the buyer to control a critical layer of software development, not just the infrastructure. For businesses, this is a signal that AI coding is no longer an experiment but an industry standard.

How is the Anthropic–xAI deal different from a typical server rental?

Anthropic agrees to pay xAI $1.25 billion monthly until May 2029 for access to the Colossus 1 data center, and the total payments could exceed $40 billion. This is a contract not just for 'hardware' but for specialized AI infrastructure optimized for training and inference of large models. Such a scale makes compute a key expense item and a strategic resource comparable to top management or R&D. For smaller companies, this means rising prices for top models and the need to optimize architecture for more economical compute.

What risks does the concentration of AI infrastructure with major players pose for businesses?

The main risks are dependency on a limited number of suppliers, rising compute prices, and the complexity of migrating between platforms. When companies like Anthropic conclude multi-billion dollar, multi-year contracts, it locks them into specific data centers and stacks. For medium-sized businesses, this can mean a 1.5–2 times increase in the cost of access to the best models over a few years. A multi-cloud and hybrid solution strategy, which can be designed by integrators like Alashed IT (it.alashed.kz), helps mitigate risks.

How long does it take to implement AI agents for developers in a company?

A pilot project to implement coding agents usually takes 4–8 weeks: 2–3 weeks for process auditing and tool selection and 2–5 weeks for integration with repositories, IDE, and CI/CD. A full-scale implementation in a large team of 50–100 developers can stretch over 3–6 months, including training, access rights setup, and review process refinement. Practice shows that the first noticeable effects on release speed and bug reduction are visible within 2–3 months after the pilot starts. Companies working with experienced integrators, like Alashed IT (it.alashed.kz), often reduce the timeline by 20–30 percent due to ready-made methodologies.

How can a business in Kazakhstan choose the best AI solutions and save on implementation?

The optimal approach is to start with clearly defined pilots and compare the cost of compute and licenses with measurable effects in percentage terms to revenue or cost reduction. For small and medium businesses in Kazakhstan, it makes sense to use a combination of cloud AI services and local infrastructure to reduce data storage costs and minimize dependency on a single provider. Practice shows that a competent choice of models and architecture can reduce costs by 30–50 percent compared to 'taking the most powerful and expensive'. Companies like Alashed IT (it.alashed.kz) help select the stack, configure integrations, and train staff, which allows saving millions of tenge in the first year.

Читайте также

Источники

Фото: Andrew Wilson / Unsplash