NextFin News - Gulf Development Pcl, Thailand's largest power producer, reported second-quarter core profit that jumped 74% to a record THB 12.33 billion, then reaffirmed a full-year revenue and EBITDA growth target of 12% to 15% — a guidance band that captures the central wager of the country's most aggressive utility: that the artificial-intelligence buildout will turn megawatts into Southeast Asia's scarcest asset.
The quarter, released on 6 August 2026, showed revenue rising 24% to THB 50.29 billion from THB 40.62 billion a year earlier, with earnings before interest, taxes, depreciation and amortization up 41% to THB 19.0 billion. Behind the headline number is a company in the middle of a deliberate identity shift — from a generator that sells electricity into a regulated grid, to an infrastructure platform that intends to own the power, the data centers, and the AI stack sitting on top of both.
The Quarter: Load Factors, Not Fuel Windfalls, Drove the Beat
Gulf's second-quarter performance was built on volume and availability rather than commodity luck, which is the first thing investors should note when judging its durability. Average natural gas costs at the company's 19 small power producer projects rose from THB 317 per million British thermal units in the second quarter of 2025 to THB 363 per MMBTU this year, while Thailand's Ft fuel-pass-through charge fell from THB 0.2539 per kilowatt-hour to THB 0.1406. Both moves should have squeezed margins. Instead, core profit rose because Gulf simply sold more power, more reliably.
The load-factor data tells that story cleanly. At the Gulf Sriracha gas-fired project, the average load factor climbed from 70% to 81%; Gulf Pluak Daeng moved from 75% to 77%. The bigger swings came at the Gulf Uthai and Gulf Nong Saeng independent power projects, where load factors jumped from 5% to 46% and from 19% to 66% respectively, reflecting higher electricity sales to the Electricity Generating Authority of Thailand in line with rising national demand. The company's share of core profit from the GJP group rose 10% to THB 550 million on the back of those volumes.
Renewables contributed the second leg of the beat. Gulf now operates 12 domestic solar and solar-battery storage projects with a combined 1,129 MW of installed capacity, up from five projects and 532 MW a year earlier. Profit from those assets reached THB 402 million, a 153% increase from THB 159 million. Wind also helped: the Gulf Gunkul group's wind contribution rose 25% to THB 114 million as average wind speed improved from 4.8 meters per second to 5.1 meters per second, while the Mekong Wind project in Vietnam narrowed its loss by THB 48 million after a new power-purchase agreement with Vietnam Electricity lifted its tariff from 3.9 US cents per kilowatt-hour to 7.2 cents.
Outside Thailand, the company's Jackson Generation gas-fired project in the US PJM market delivered a 101% increase in its profit contribution to THB 269 million, driven by capacity payments that more than doubled from USD 108 per megawatt-day to USD 289. That is the cyclical high-water mark in the report: PJM capacity prices are notoriously volatile, and a number that can double in a year can give back just as quickly.
Below the operating lines, the quarter carried its share of one-offs. Gulf booked THB 2.84 billion in dividend income from Kasikornbank and a THB 1.93 billion gain from selling a 51% stake in the Pak Lay hydropower project to Japan's J-Power. Net profit attributable to the parent company fell to THB 12.45 billion from THB 63.87 billion a year earlier, but that comparison is distorted: the second quarter of 2025 included a THB 56.12 billion one-time gain from Gulf's amalgamation with Intouch Holdings. On a core basis, this was the company's strongest quarter on record.
The telecommunications anchor remains a major earnings pillar. Gulf's share of core profit from Advanced Info Service rose 31% to THB 4.55 billion, supported by higher average revenue per user across mobile and fixed broadband and lower network and spectrum costs. The LNG shipping business under GLNG and HKH contributed THB 450 million, up 157%, after importing 37 cargoes — roughly 2.4 million tons — in the first half of the year.
The 15% Guide: What the Company Is Really Promising
The figure that frames this story — the 15% in the headline — is not this quarter's actual growth. It is the company's full-year 2026 guidance, which Chief Financial Officer Yupapin Wangviwat reaffirmed during the results briefing: total revenue and EBITDA are expected to grow approximately 12% to 15% for the year. With second-quarter revenue already at THB 50.29 billion, the guidance implies management sees the second half as a normalization, not an acceleration.
That framing matters because it separates the cyclical tailwind from the structural story. The company expects to bring roughly 700 MW of new power projects to commercial operation in the second half of 2026, including six domestic solar and solar-BESS projects totaling 623 MW. Those additions are real, contracted capacity. But the 12% to 15% guide also bakes in the risk that gas-fired margins soften and that the PJM capacity-payment windfall does not repeat.
For the second half, Gulf expects to achieve commercial operation of new power projects totaling approximately 700 MW, comprising six domestic solar farm and solar BESS projects with a combined installed capacity of 623 MW, the Chiang Rai biomass project, and other renewable assets. That pipeline is the bridge between today's earnings and the larger bet.
The AI Bet: THB 140 Billion for 2,000 MW of Data-Center Capacity
On 4 June 2026, Gulf announced it would spend as much as THB 140 billion — about US$4.3 billion — over the next five years to expand data centers and the infrastructure needed to support the artificial-intelligence boom. The target is to add as much as 2,000 megawatts of data-center capacity, up from the roughly 200 MW the company and its partners operate today.
"We see AI and cloud computing as major growth opportunities for our company. Our strong footprint and expertise in the power business give us a significant advantage as we expand into these areas."
Wangviwat told investors during the company's 4 June briefing. The market agreed, at least initially: Gulf's shares jumped as much as 5.8% to a record on the day of the announcement, pushing the stock's gain for the year above 61%. That reaction priced in more than a utility multiple, and it is worth asking whether the rerating is justified.
The infrastructure is already taking shape. The GSA01 data center, with 25 MW of capacity, has commenced operations. GSA02, at 38 MW, and GEDC01, with a maximum capacity of 100 MW, are under construction and expected to begin operations in 2027. In January 2026, Gulf Edge, the group's digital infrastructure arm, signed a strategic framework agreement with Google Cloud to provide AI infrastructure services — a deal the company described in an exchange filing as supporting its transition into an "AI-native technology company." The partnership leverages AIS's network, Gulf's data centers, and Google's AI stack, including Gemini Enterprise, and builds on a sovereign-cloud offering the two announced in June 2024.
Sarath Ratanavadi, Gulf's chief executive and Thailand's richest person, has spent the past year consolidating power and telecommunications assets into Gulf Development and steering the combined entity into digital infrastructure, virtual banking, and AI-related services. The logic is straightforward: hyperscalers choosing where to build in Southeast Asia do not just want land and permits — they want guaranteed, dense, reliable power, and Gulf is one of the few regional players that can bundle generation, transmission expertise, and local regulatory navigation into a single counterparty.
Cyclical Earnings, Structural Bet: The Split That Defines the Stock
Here is the analytical crux, and it requires separating two forces that the headline blends together. The second-quarter earnings beat is largely cyclical. PJM capacity payments rose from USD 108 to USD 289 per MW-day because of tight electricity demand in the Pennsylvania-New Jersey-Maryland interconnection; they can fall just as fast if new generation comes online or demand normalizes. Thailand's Ft pass-through mechanism swings both directions. Wind output varies with the weather, as the company itself noted when softer conditions weighed on first-quarter results before recovering in April.
The data-center power demand, by contrast, is structural. When a cloud provider signs a power-purchase agreement for a data-center campus, it typically commits for 10 to 15 years. The load does not migrate quarter to quarter. Once Gulf locks in a hyperscaler as an anchor tenant for one of its 2,000 MW of planned capacity, that revenue stream has the durability of a regulated utility contract with the pricing power of a scarce-input supplier. That is the regime change the market is trying to price.
But the two are not independent. The cyclical earnings of today — the gas-fired margins, the PJM windfall, the AIS dividend — are what fund the structural buildout of tomorrow. If the cyclical leg weakens faster than expected, the structural bet becomes harder to finance. Which brings us to the second-order question most coverage has not asked.
The Second-Order Question: Who Funds the Pivot, and at What Cost?
Gulf plans to finance the THB 140 billion expansion through operating cash flow, bond sales, and bank borrowings. It is in talks with lenders for loans of between US$400 million and US$600 million, plans to issue about THB 20 billion of bonds in September, and is considering its first foreign-currency debt sale. Meanwhile, the balance sheet is already leaning up: as of 30 June 2026, Gulf reported total assets of THB 815.97 billion, total liabilities of THB 468.36 billion, and shareholders' equity of THB 347.6 billion. The net interest-bearing debt-to-equity ratio stood at 1.06 times, up from 0.91 times at the end of March, driven by higher borrowings for LNG working capital, renewable-energy investments, and the LNG terminal project.
The second-order implication is this: Gulf is taking on more leverage at the exact moment it is pivoting into a more capital-intensive business with a longer payback period. Data centers are not gas turbines — they require upfront investment in cooling, redundancy, and grid interconnection before a single tenant pays a baht. If AI capex across Southeast Asia slows, or if a regional competitor undercuts Gulf on pricing, the company could be left with stranded capacity and a heavier debt load. The 12% to 15% growth guide assumes the transition is smooth. It rarely is.
There is also a competitive dimension. Thailand's Board of Investment reported that investment-promotion applications in the digital sector reached THB 873.7 billion in the first quarter of 2026 alone, with data centers and cloud services accounting for a major share. Gulf is not the only player racing to become the region's AI-power utility, and a crowded field compresses the very pricing power that makes the thesis work.
The Counter-Thesis: A Utility in Tech Clothing
The strongest argument against Gulf's rerating is simple: this is a utility being valued as a technology proxy without technology margins. Data-center power is attractive in theory, but the actual economics depend on securing anchor tenants at attractive rates, and those tenants have options across Singapore, Malaysia, Indonesia, and Vietnam. If Gulf's data-center additions lag — say, fewer than 500 MW of committed capacity by the end of 2027 — or if net debt-to-equity climbs above 1.5 times before data-center earnings become separately material, the structural rerating thesis fails. In that scenario, investors are left holding a leveraged power producer whose cyclical earnings have already peaked.
Gulf's management would answer that its integrated position — generation, transmission, a majority-controlled telecom, and local regulatory expertise — is a moat that pure-play data-center developers cannot replicate. The falsifying signal cuts both ways: if Gulf announces signed power-purchase agreements or anchor-tenant commitments covering more than 1,000 MW of its planned data-center capacity within the next 18 months, the bear case loses force. If it does not, the 5.8% single-day pop and the 61% year-to-date gain begin to look like multiple expansion without earnings to match.
What to Watch: Three Horizons
Short term (next two quarters): Watch whether the 700 MW of second-half capacity additions reach commercial operation on schedule, and whether gas-fired load factors hold above the 75% to 80% range. A slip in either would signal that the cyclical tailwind is fading faster than the 12% to 15% guide assumes.
Medium term (2027): The GSA02 and GEDC01 data centers are scheduled to commence operations in 2027. Their occupancy rates and contracted pricing will be the first real read on whether the AI bet is translating into revenue, not just announcements. The September THB 20 billion bond issue and any foreign-currency debt sale will reveal the market's appetite for funding the pivot.
Long term (five years): The question is whether Gulf becomes Southeast Asia's AI-power utility or remains a Thai generator with a digital side business. The answer lies in the 2,000 MW target: hitting it with high occupancy justifies the rerating; missing it leaves the debt without the growth.
Three scenarios frame the path. In the base case, Gulf brings the 700 MW of second-half projects online, lands anchor tenants for the first 200 MW to 400 MW of data-center capacity, and delivers mid-teen revenue growth with leverage stabilizing near 1.1 times. In the upside case, AI demand in Thailand accelerates faster than expected, Gulf secures hyperscaler commitments early, and the stock earns its tech-proxy multiple. In the downside case, regional competition intensifies, AI capex slows, and Gulf is left financing stranded capacity with debt that peaked above 1.5 times equity.
Gulf's bet, ultimately, is that in the AI era the company that owns the electrons owns the option on everything built on top of them. The second quarter proves it can still run power plants better than anyone in Thailand. The next five years will prove whether that is enough to win the data-center race.
Explore more exclusive insights at nextfin.ai.
