India’s AI Inference Problem, Anscer Bags ₹45 Cr & More

The Dollar Cost Of AI Inference

India’s AI boom may be quietly creating its next major dollar outflow problem.

As startups and enterprises rush to embed generative AI and large language models into everything from customer support to internal workflows, a growing share of their technology spending is flowing overseas through AI inference bills.

The New AI Bill: AI inference is becoming the biggest operating cost of AI-native products. Unlike classic software, every chatbot reply, workflow trigger or personalisation step can create a fresh bill, often tied to tokens or cloud compute. That means AI usage rises not just with customers, but with each interaction.

The Dollar Dent: For many early stage startups, AI inference already devours a big chunk of their cloud infrastructure budgets, a figure expected to rise even more going forward. With foundational models and GPU capacities tethered to dollar-linked pricing, Indian companies find themselves earning in Rupees but paying for their critical AI operations in dollars, which are at the mercy of market fluctuations.

The Value Leak: The concern is not just cost inflation, but also where the money flows. Much of India’s AI consumption ends up enriching overseas model providers and infrastructure owners rather than domestic compute ecosystems. As a result, critics are flagging that AI compute costs are starting to resemble a digital version of the import bill.

Engineering A Local Pivot: To counteract this economic drain, domestic players like Yotta, CtrlS and Reliance Jio are expanding GPU clusters and AI cloud offerings. Industry executives believe India has structural advantages in this space like a thriving developer ecosystem, domestic demand and lower operating costs for LLM makers.

Simultaneously, enterprises are already testing the local infrastructure for lower latency, better data residency and tighter control over costs. With the country projected to guzzle around 7 GW of AI compute capacity by 2030, will India remain dependent on foreign AI compute or will the domestic infrastructure step up? Let’s find out… 

From The Editor’s Desk

Anscer Robotics Bags ₹45 Cr 

• The industrial robotics startup has raised $4.6 Mn in its Series A round led by IAN Alpha Fund to strengthen its product stack, scale US market operations and expand global partner network.
• Founded in 2020, Anscer Robotics builds autonomous mobile robots, fleet management software, and industrial automation systems for factories. It has raised $7 Mn to date and operates a facility in Bengaluru that can manufacture 1,000 robots annually. 
• As warehousing operations grow more intricate, businesses are automating operations to improve efficiency. On the back of this, the homegrown industrial automation market size is projected to cross $25.8 Bn by 2028. 

Ola Electric’s Revenue Dips

• The EV maker managed to trim its consolidated net loss by 43% YoY to ₹500 Cr in Q4 FY26. Operating revenue also slumped 57% YoY to ₹265 Cr during the quarter as E2W deliveries shrank 61% YoY to 20,256 units.
• However, total expenses for the quarter declined 57% YoY to ₹546 Cr. As a result, the company claimed that Q4 was its first operating cash flow positive quarter. 
• Going forward, the EV maker said that it is targeting a recovery in national market share to 15% to 20% over the next six months through better service, stronger store productivity and new product launches.

upGrad Nears Profitability

• The edtech major claims to have turned PAT profitable in the first 11 months of FY26. upGrad is said to have posted a net profit of ₹38 Cr on provisional and unaudited revenue of ₹1,532 Cr during the period. 
• The unicorn is expected to close FY26 with revenue of ₹1,972 Cr and PAT exceeding ₹60 Cr, against a net loss of ₹274 Cr in FY25. Despite the boost, the company’s valuation slipped to $1.7 Bn, down 24% from its peak valuation of $2.25 Bn in 2022.
• This comes as upGrad is all set to acquire rival Unacademy in an all-stock deal, which would peg Unacademy at around ₹2,055 Cr (about $218 Mn). This implies a 90% cut from the troubled edtech’s peak valuation of $3.4 Bn in 2021.

Inside ixigo’s AI Push

• For more than a decade, travel apps have barely changed. Users still have to juggle tabs, compare fine print and worry about everything from delays to visa rules. ixigo wants to change this by using AI to turn travel planning into a conversational experience. 
• The OTA’s AI assistant, TARA, is designed to do the work users usually do manually, from finding shorter layovers to checking hotel prices. The company also claims to have processed over 1.5 Lakh calls end-to-end through AI systems during a disruption.
• In the long run, ixigo wants to make AI the glue between product, support and retention, so that users start seeing the platform not as a place to book tickets, but as a reliable travel companion they can carry in their pocket.

Lenskart’s Q4 Show

• The eyewear giant’s consolidated net profit declined 7.5% YoY to ₹203.6 Cr in Q4 FY26, despite expanding margins and operating revenue zooming 46% YoY to ₹2,515.7 Cr. What played a spoilsport was expenses surging 36% YoY to ₹2,308.3 Cr.
• For the full FY26, Lenskart’s profit surged 69% YoY to ₹500.9 Cr, aided by strong operating leverage, premiumisation, and international growth. Meanwhile, its top line rose 33% to ₹8,814 Cr during the fiscal.
• The company added 183 net new stores during the quarter in India, taking the total store count to 3,327 globally. In FY27, the company plans to embed AI across manufacturing, customer experience, supply chain and store operations to drive operating leverage. 

Inc42 Markets

Inc42 Startup Spotlight

Can Pramatra Safeguard India’s Quantum Stack?

Quantum computing could eventually break today’s encryption. However, most current systems are not ready for that shift as existing quantum-safe options, especially fibre-based ones, remain limited by range and scalability. Enter Pramatra Space, a startup trying to fix this problem.

Reimagining Security: Founded in 2023, Pramatra Space is developing satellite-based Quantum Key Distribution (QKD) to create encryption keys that can be transmitted securely across long distances. By moving the problem into orbit, the startup aims to overcome the physical limits of terrestrial fibre networks and make global communication secure.

Chips In Orbit: The startup’s approach uses integrated photonic chips mounted on low Earth orbit (LEO) satellites to generate and transmit quantum keys to ground stations. Those keys are designed to be extremely difficult to intercept, giving enterprises and critical infrastructure a new layer of protection against future cyber threats.

Beyond Key Exchange: Pramatra is also planning to offer key management systems to secure digital assets and mission-critical networks. Alongside that, it is developing specialised chips for both satellite payloads and ground-based security modules, which could help it control more of the security stack and improve deployment flexibility.

As both governments and enterprises prepare for a post-quantum future, the startup is eyeing a piece of the global quantum cryptography market, which is projected to cross $4.6 Bn by 2030. So, can Pramatra build a secure backbone for India’s digital communication?

Infographic Of The Day

India’s watch market is no longer just Titan and Timex. From D2C microbrands to premium curators, a whole new generation of Indian watch brands is quietly building a market that didn’t exist a decade ago…

The post India’s AI Inference Problem, Anscer Bags ₹45 Cr & More appeared first on Inc42 Media.